Summary
Highlights
The video introduces the discussion on visual pathway organization beyond the primary visual cortex (V1), examining the flow of information to various brain areas, specifically two major pathways. It also briefly touches upon the fate of ganglion cell axons not following the predominant visual pathway.
The dorsal stream, known as the 'where' pathway, processes information about an object's movement in space and guides body movements to interact with objects. Information from V1 passes through secondary and tertiary visual cortices, targeting area MT (motion perception), then proceeds to the posterior parietal lobe for integration with other senses, and finally to motor control areas for planning visually guided movements.
The ventral stream, or 'what' pathway, processes information about an object's features, including form and color, which originate from V4. This pathway involves areas along the inferotemporal cortex, maintaining detailed representations of familiar stimuli.
The secondary visual cortex (V2) segregates information for the dorsal and ventral pathways. V2 consists of bands that receive projections from different V1 areas, representing distinct stimulus features. Thin stripes encode color information (from V1 blobs) and project to the ventral pathway, while interstripes (from V1 interblob regions) encode form information and also project to the ventral pathway. Thick stripes (from V1 input layer) pass motion information to area MT in the dorsal pathway.
Evidence for these two pathways comes from primate experiments. Monkeys with parietal cortex lesions (dorsal pathway) failed a landmark task (location-based), while monkeys with inferotemporal cortex lesions (ventral pathway) failed an object recognition task. This 'double dissociation' indicates separate processing channels for 'where' and 'what' information.
Similar results are seen in humans with brain damage. Patient DF, with bilateral ventral stream lesions, suffered from object agnosia (unable to identify objects) but could still use visual information to guide actions, implying an intact dorsal stream.
The inferotemporal cortex has specialized areas for different objects of visual experience, such as the parahippocampal place area (environmental scenes), LOC (familiar objects), a body area, and the fusiform face area (FFA). The FFA was initially thought to be face-specific but research suggests it responds to objects with which individuals have significant expertise.
The video discusses how the brain encodes faces and other familiar stimuli. Instead of a 'specificity code' where individual neurons encode specific faces, researchers believe in a 'distributed neural code.' Here, multiple neurons respond to various faces, and the collective pattern of their output uniquely identifies each individual face or highly familiar object.
About 10% of retinal ganglion cell axons diverge from the main pathway and target the superior colliculus in the midbrain. This structure integrates visual, auditory, and tactile inputs to orient the eyes towards stimuli of interest, like potential threats or rewarding objects, leveraging the visual system's spatial abilities.
Tracking visual stimuli involves two processes: smooth pursuit (gradual eye movements to minimize retinal offset for slow-moving objects) and saccades (rapid eye movements for large, quick object movements). While historically seen as distinct, current models suggest highly integrated pathways for these movements, reflecting their natural blend in tracking.